Re:Depressed woman loses benefits over Facebook photos
by pyre on Thursday November 19, 2009 @10:12PM
(#951800)
While that may be the case, in this instance the insurance company is probably in the wrong. The Facebook photos along should not be enough for them to say that she's cured. Maybe enough for them to feel they need to spawn an investigation, but I somehow doubt that they consulted with her doctor or had a doctor of their own choosing interview her.
An insurance claims agent looking at Facebook photos is not enough to say, "Look! There's a smile on her face! She can't be depressed! It's impossible for a depressed person to smile!" This isn't exactly the same as seeing a photo of someone that claims to be paralyzed from the waist-down running in a marathon...
How the insurance company accessed the photos in this instance is irrelevant. I'm more concerned with how they used them once they had them. I'd be concerned with how they obtained them if there was a possibility of using 'dubious means' though.
Suppose Facebook applies some Machine Learning algorithm and can predict roughly which users will incur insurance claims. Inputs could include user-generated text, size and shape of social graph, entities recognized in user-generated photos, etc. That data would be valuable to insurers.
It seems like many of these users are mentally in a living room with their friends, while actually uploading their private info to a corporation's servers.